#linuxcnc-devel Logs

Nov 04 2017

#linuxcnc-devel Calendar

03:36 AM tjtr33: seb- still trying, i have gotten orig rip of dynamic offsets to hang axis while zooming halscope.
03:38 AM tjtr33: maybe my video dunno. once axis hangs i dont know how to clean.
03:38 AM tjtr33: i get 'lockfile' and cant remove lock in /tmp succesfully. so testing is slow. i i can recreate consistantly will post ug report.
03:40 AM tjtr33: i went there becasue i saw xy values from siggen seemed wrong. there is a lag between G:waveform and E:z-offset that accounts for circles becoming squircles ( imo )
03:41 AM tjtr33: thats the reason for the acc and vel spinboxes, if they are raised quite a bit, the flattened sinus matches the true sinus and the lag reduces.
03:42 AM tjtr33: ttul8r
03:57 AM tjtr33: well thats pretty easy to scrag Axis,just view those 2 signals in halscope, then move the axis window. you are now frozen.
03:58 AM tjtr33: now to learn about submitting bug report. willleave frozen axi window for a couple hours to see if it unzombiefies
04:04 AM tjtr33: oh and the output 'E:z-offset' is a rounded triangle, not a sinus.and lags many degrees, hard to measure
04:10 AM linuxcnc-build: build #363 of 1540.rip-jessie-armhf is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1540.rip-jessie-armhf/builds/363 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
04:58 AM linuxcnc-build: build #5179 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/5179 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
08:50 AM JT-Plasma: I'm getting a permission denied when I try and loadrt hm2_pci http://paste.ubuntu.com/25886421/
09:32 AM JT-Shop: never mind I forgot to do a sudo make setuid
09:52 AM JT-Shop: Dewey if you happen to drop by I have two projects today dgarr/external-offsets and some grass boxes for the chickens
10:59 AM tjtr33: JT-Shop could you the external offsets demo ?, it's a sim that jiggles Z with a sine generator.
11:00 AM tjtr33: my system hangs if i adjust the hals scope and i dont want to file a bug report if its just my system.
11:00 AM tjtr33: its very repeatable here
11:02 AM pcw_home: tjtr33 anything in dmesg (in a terminal window)?
11:02 AM tjtr33: axis will hang, yoiu will notice a 'glitch' a temporary whatsthat when looking at halscope and then notice axis is hing, the screen shreds when moved
11:02 AM tjtr33: reooting now bt i have some captures that dont suggest anything
11:02 AM tjtr33: i hae some dmesg captures that dont suggest anything
11:02 AM pcw_home: sounds like a video bug (most will have a dmesg indication of trouble)
11:03 AM tjtr33: yes i suspect video but see nada in dmesg and logs
11:03 AM tjtr33: any special place to look? btw othe apps run fine but lcnc cant be run again ( lockfile )
11:04 AM tjtr33: halscope keeps going too
11:04 AM pcw_home: OK the problems I've seen hang all video, though you can eventually get a virtual terminal
11:05 AM tjtr33: is it valid to run a different sim and play with halscope as a debug measure?
11:06 AM tjtr33: its easy to get another termonal running, but for 2 days now ive never cleaned lcnc enuf to run again, nor had it heal itself
11:07 AM tjtr33: i have to reboot after a glitch
11:08 AM tjtr33: rebooted and testing again now
11:14 AM JT-Shop: tjtr33: I had a problem with my system hanging last time I messed with it but it's been quite a while since I last messed with it
11:16 AM tjtr33: thx jt testing again now
11:16 AM JT-Shop: which config are you running?
11:17 AM tjtr33: dynamic_offsets
11:18 AM tjtr33: i just rememberd i wanted to reverify the latency tests before testing the hang again
11:18 AM JT-Shop: by adjust do you mean change size or position of the window?
11:21 AM tjtr33: adjust the pyvcp vel from 2 to 1, then adjust the scope so the 3 waveforms are about the same size, seperate them so each is nicely visible, then youll see this tiny pause in the halscope update and notice that if you move the axis window, it tears
11:22 AM dgarr: tjtr33: i can't reproduce your problem, do you have both a rip build and a deb install on the system?
11:22 AM tjtr33: yes, external offsets is a rip
11:23 AM dgarr: do you also have a deb install on the same system is my questioin
11:24 AM tjtr33: hmm i installed from a linuxcnc iso, then built the rip
11:24 AM tjtr33: i never used a .deb to install lcnc, ity is from an iso installtion image
11:25 AM dgarr: having both a deb install and a rip can cause problems unless you are very careful to source the rip-enviromnent script in every shell that uses the rip build
11:25 AM tjtr33: . scripts/rip<tab> litany?
11:25 AM dgarr: if you installed from iso, then you probably have a deb install (unless you removed it)
11:25 AM JT-Shop: tjtr33: I can't reproduce it
11:26 AM dgarr: show the output of the command: $ dpkg -l "linuxcnc*"
11:26 AM JT-Shop: dgarr: hi, I'm giving up on work today :) and working on external offsets
11:26 AM tjtr33: on another system for inet
11:27 AM tjtr33: Desired=Unknown/Install/Remove/Purge/Hold
11:28 AM tjtr33: | Status=Not/Inst...bblahblah'un linuxcnc <none> <none> (no description available)
11:29 AM tjtr33: i can run linuxcnc from a terminal WITHOUT the . scripts/rip<tab> and it says LINUXCNC - 2.7.6
11:30 AM dgarr: the 2.7.6 probably means you are running the deb install -- learn how to use the unix command which, for example:
11:30 AM tjtr33: and WITH the . scripts<blah> it says LINUXCNC - 2.8.0~pre1 and has the offsets configs available
11:30 AM dgarr: $ which linuxcnc # will show the executaable
11:30 AM tjtr33: read what i said.
11:30 AM tjtr33: please
11:31 AM dgarr: i understand-- you hae two installs, one is a deb and one is a rip, when using the rip you must source the rip-environment script in
11:31 AM dgarr: every shell.
11:31 AM dgarr: i recommend for the least confusion that you remove the deb install if you are testing a rip
11:32 AM tjtr33: yes i am sourcing when i have the problem, i have no trouble running rips.
11:32 AM dgarr: i recommend that you remove the deb install
11:33 AM tjtr33: remove the one that works so i can test the one that doesnt?
11:33 AM dgarr: yes
11:33 AM tjtr33: ok thanks for the advise
11:33 AM dgarr: sudo dpkg -r name_of_the_installed_deb
11:34 AM tjtr33: it was not installed from a deb. it was installed with an iso image on a thum drbe
11:34 AM dgarr: yes that means it was installed from a deb
11:35 AM tjtr33: ok
11:35 AM dgarr: the command dpkg -i "linuxcnc*" will show the name ofthe installed deb, the command sudo dpkg -r name_of_deb will remove it
11:37 AM tjtr33: sudo dpkg -i "linuxcnc*" -> dpkg error processing archive: no such file or directory
11:38 AM dgarr: i mis typed, sorry dpkg -l "linuxcnc*" will list packages (not -i , -i is for install)
11:38 AM tjtr33: Applications menu | CNC | linuxcnc runs fine
11:38 AM tjtr33: ok
11:40 AM tjtr33: linuxcnc , linuxcnc-doc linuxcnc-sim linuxcnc-uspac
11:40 AM tjtr33: ok but i am leary of removing the working for the non-working to debug something i have been told is a video problem
11:41 AM dgarr: if you think it is a video problem you could verify with other sims and not cast blame on external offsets
11:42 AM tjtr33: i have been askign what to suspect, and casting blame on my attempts not you.
11:43 AM tjtr33: others suggest it is a video problem, thats my first advice since first asking 2 days ago
11:43 AM tjtr33: ok thx for your opine.
11:47 AM dgarr: when one has both a deb and a rip install, one has to be very careful with the shell environment because shell commands executed within linuxcnc sripts use the unix PATH to find executables -- it is often simplest to make sure there is only one resolution to PATH searches.
11:48 AM dgarr: removing a deb makes rip testing less error prone, it is easy to reinstall production (or buildbot debs) when necessary
11:56 AM dgarr: JT-Shop: i suggest you experiment with configs/sim/axis/external_offsets/eoffset/hpid.ini or hpid_mm.ini and study the following for adapting to hardware -- good luck
11:56 AM dgarr: https://github.com/LinuxCNC/linuxcnc/blob/dgarr/external_offsets/docs/src/motion/external-offsets.txt
11:56 AM dgarr: and especially: https://github.com/LinuxCNC/linuxcnc/blob/dgarr/external_offsets/configs/sim/axis/external_offsets/eoffset_pid/hpid.txt
11:59 AM JT-Shop: has anything changed in the two txt files lately? I printed them a few months ago
11:59 AM JT-Shop: I'm up to a working configuration just need to tune it I think
11:59 AM dgarr: from months ago -- yes, please read again
12:01 PM JT-Shop: ok, I'll print them out and toss the old ones
12:02 PM dgarr: the sim configs allow you to experiment with tuning -- you can pause with the simulations but pausing is not likely practical with hardware torch testing. I suggest you adapt a sim with the exact velocity, accel, and position limits and use the perturbation facilities to examine effects of the many items that can affect response of the loop
12:04 PM dgarr: crashes in a sim due to loop instability is better than crashing a torch i imagine
12:04 PM JT-Shop: I have a floating head so no damage if it touches the work
12:06 PM dgarr: there are many variables to set, so using a sim with replicated ini items may reduce the number of attempts with hardware.
12:13 PM JT-Shop: ok, thanks
12:15 PM dgarr: JT-Shop: also, you probably should make sure your build is uptodate since i force push that branch, the current commit is 7affcf2
12:16 PM dgarr: git log --oneline -1 origin/dgarr/external_offsets
12:16 PM dgarr: 7affcf2 External Offsets: exclusive dir for eoffset_pid
12:18 PM JT-Shop: I did a fresh clone this morning
12:18 PM dgarr: ok, that should be 7affcf2
12:19 PM dgarr: (but you should make sure)
12:23 PM JT-Shop: yes I'm at 7affcf2, but git log didn't like origin/dgarr/external-offsets but did work with just dgarr/external_offsets
12:25 PM dgarr: external_offsets not with hypen (external-offsets) maybe? i think 'origin' is defined in the git config files ( a hidden top dir named .git)
12:27 PM JT-Shop: yea just looked and I didn't use the underscore, have to look here the slid over there and type :)
12:28 PM JT-Shop: I have to run to town and fetch some grub fixins be back a bit later
02:24 PM dgarr: tjtr33: experimented with two-axis external offsets and could not reproduce problems: https://youtu.be/pzXRYoG1UPg
02:24 PM dgarr: the complete config is at: http://www.panix.com/~dgarrett/stuff/xy_offsets.tgz
02:24 PM dgarr: a) control the ramp-up/ramp-down of offset amplitudes to avoid step changes,
02:25 PM dgarr: b) coordinate offset amplitude & frequency in conjunction with gcode feed rates and the machine's [AXIS_*]MAX_VELOCITY,MAX_ACCELERATION,OFFSET_AV_RATIO settings
02:25 PM dgarr: c) handle cutter radius compensation according to independently set (and possibly dynamic) offset magnitude.
02:25 PM dgarr: An 'all-gcode' solution seems more suitable to me.
02:26 PM dgarr: left out in paste: For trochoidal milling, external_offsets adds numerous *knobs* that would need
02:26 PM dgarr: adjusting to: a),b),c):
02:36 PM IchGucksLive: hi Nice Feature woudt also bee to set the Tool with a comment G10 L1 P1 Z0 R5 ;endmill R5
02:37 PM IchGucksLive: as lots of like to read what tool it is
02:37 PM IchGucksLive: its not a easy task
02:37 PM IchGucksLive: as not numeric
02:38 PM IchGucksLive: but maybe a list of numerics that printout endmill /drill .....
03:09 PM hazzy: What is the best way to get a list of HAL pins in python?
03:09 PM hazzy: Right now I am doing this: subprocess.check_output(['halcmd', 'show', 'pin'])
03:09 PM hazzy: Is there a more direct approach?